Negative Peak Limiting Circuit Single diode

[W7XXX]

By request here is the circuit I have used with success since 1999. Excuse the crude sketch and writing!

[W7XXX]

The writing is impossible to read so here is the scoup.

The resistance of the resistor should be about double your final resistance for starters.
For 125% mine is adjusted at 12.5k now and is 225 watts for a single 813. For my former 100 watt out rig, I used 3 times the resistance as the goal was to modulate only 105% postive, prevent splatter, and get negative peak up close to 100% and also protect the modulation xfmr.

The wattage should be at least half the RF output. Overkill is in order here.

You must also modulate the screen grid with the plate.

Be sure your final can handle the increased voltage (AC component).

Use a scope to adjust and always use it to monitor your signal.

PEP is figured by carrier power times 4 ... times percent positive modulation.
Example: 300 watts carrier x 4 = 1200 watts x 1.25 (125% +) = 1500 watts PEP.
This is how the FCC figures it whether we like it or not.

Only big diodes I could find in 1999 were the K2AW. They have never failed me.

[Pete, WA2CWA]

See if the picture works better now. Open it with Microsoft Image Composer or something similar. Enlarge to something greater than 100%.

[k4kyv]

A problem I have with what I see is with the negative peaks.  The envelope pattern as drawn looks like something  from the upside-down tube circuit or a high-level balanced modulator.  A conventional plate or grid modulated rf final will simply show a straight line at the negative peak of the audio waveform (aka overmodulation), as the rf output from the final momentarily drops to zero; it will not produce the "dog-knot" as shown in the drawing, which indicates a phase-reversal and continued rf output as the final amplifier plate voltage swings into the negative, continuing below the base-line.

This scheme, known as "negative cycle loading" can be expected to cause at least as much splatter as the overmodulation it is designed to prevent. On the negative half of the audio cycle, as seen at the secondary of the modulation transformer, the diode conducts and places the resistor in parallel with the class-C rf final as a load on the modulation transformer.  If the modulator tubes are looking into the correct plate-to-plate load impedance as calculated by the class-C modulating impedance (rf final plate voltage/plate current) and  modulation transformer turns ratio, then there will be an impedance mismatch during the entire negative half-cycle of the audio because of the lower impedance load presented to the modulator tubes by the parallel resistance.  This mismatch itself may not be a reason for additional splatter, but it is the abrupt change in load impedance as the output voltage swings past the base line that will cause the exact same kind of splatter due to higher order harmonics as plain and simple overmodulation in the negative direction.

A variation of this circuit can be used to protect the modulation transformer (and also serve as an overmodulation indicator):  change the resistor to a value equal to the class-C final amplifier modulating impedance (Vp/Ip), and return the resistor as shown to ground instead of to B+.  That way, if the final amplifier is overmodulated by driving the plate voltage into the negative, the diode conducts and places the resistor in the circuit to maintain the load impedance on the modulation transformer during the negative part of the cycle until the PA plate voltage returns to the positive range, and the modulation transformer is not operating over part of the audio cycle with no load, which can be death to a modulation transformer.  This scheme will not affect the modulated waveform, but will guarantee that the modulation transformer is looking into a proper load even if overmodulation in the negative direction occurs.  Use a M-V rectifier tube like the 866-A as the diode, and you have a nice overmodulation indicatior, as the tube will flash blue inside whenever the PA plate is driven negative.

Whether using single diode negative cycle loading or the three-diode 'ultramodulation' circuit, there is a fallacy to the whole concept of attenuating the negative half of the audio cycle of the modulating signal.  Firstly, part of your expensive-to-produce audio is being rectified and converted into DC which is dissipated to heat up a resistor.  The waveform, which otherwise might be a sinewave, becomes distorted.  This circuit essentially converts it into a controlled-carrier AM signal with significant even harmonic distortion added to the audio.  With ultramodulation, the measured plate current and plate voltage actually rise with modulation as measured on a DC ammeter and voltmeter.  The upward kick of the final amplifier plate current, the brightening of the colour on the PA tube plates, the extended positive peaks of the waveform that appear on the envelope pattern of the scope, and the increased upward kick of a thermocouple rf ammeter in the antenna feedline might look good and stroke the ego, but do not result in increased intelligibility or loudness of the audio.  The extra input and output power you see are composed largely of spurious  distortion products.  Precisely the same effect could be produced with conventional controlled-carrier modulation and carefully produced pre-distortion of the audio signal fed into the modulator.

There is no need to deliberately distort the audio waveform with diodes and resistors  to take advantage of extended positive peaks. First of all, the modulator must be capable of generating enough undistorted audio power to allow for positive modulation peaks beyond 100% - the rate increases as the square of the modulation percentage.  In other words, it takes four times as much audio power to modulate 200% positive as it does to modulate 100%.  So the first requirement is to install an extra beefy modulator for the power level of the final.  Secondly, make sure that the entire audio chain has as low distortion as possible.  The audio response capability of the system should be at least one octave above and below the audio frequency range you intend to transmit.

Most human voices naturally have a distinct asymmetry to the audio waveform.  So  once the two requirements as stated above are satisfied, it is a simple matter of phasing the audio so that that the peak with the greater amplitude extends in the positive direction.

I don't know what calculating p.e.p. has to do with this topic, but... well here's my opinion of p.e.p.:

http://www.youtube.com/watch?v=WxmUKVrT0iI

[Steve - WB3HUZ]

It's far easier to contorl peak, negative, positive or both at the line level portion of the audio system. Other than for some sort of ultimate mod tranny protection, I see little use for such high-level schemes.

[steve_qix]

Hi !

The single diode "loading" type of circuit assumes a modulator with plenty of internal resistance (loss).  The perfect modulator is an ideal voltage source - being a source which will deliver the requested voltage into any load.  The voltage output is unaffected by the load (and by extension, will supply infinite current).  Of course, this is not practically achievable, however it is possible to come fairly close.

An ideal voltage source modulator is approchable by using low impedance output devices, negative feedback, and good quality transformers.  The solid state analog modulators I use with the class E rigs come VERY close to being ideal voltage sources over the design power range of the modulator.  The output does not vary between no load and a 200% of rated output (100% overload).  A very good modulator just "gives".

What this means, is that you will place undue stress on the modulator components (tubes, transformers, etc) by using the single diode-resistor peak limiter.  You will also waste a lot of power.  Using the 3 diode method does not place any additional stress on the modulator.  If you are worried about "hard edges", you could soften them somewhat by adding a resistance or inductance in series with the "keepalive" diode, but in my experience (35+ years), it is a non-issue.

The largest transmitter I've seen with a high level (3 diode) negative peak limiter was the Collins 21-E at WORC (in Worcester, MA) back on the '70s.  This was at a 5000 watt broadcast station.  I made the modification to the transmitter when I was chief engineer.  The 21-E would *easily* modulate 140% positive (which was illegal, and we did not do it), but it was interesting to see what the transmitter would do 

Regards,

Steve

[ka3zlr]

I seen that very circuit diagram before.. it was on a DX-100 site..I had my own reservations on that.

[w3jn]

Steve, you had posted a nice treatise on the 3-diode limiter (or perhaps is was on your site)?  I can't find it here for some reason.  WOuld you have a link?  That was a Fine Business article indeed.

[steve_qix]

Steve, you had posted a nice treatise on the 3-diode limiter (or perhaps is was on your site)?  I can't find it here for some reason.  WOuld you have a link?  That was a Fine Business article indeed.

Hmmm... The only thing I can *remember* (that's the key word -remember :-) :-)  was the schematic and description.   I'll post it here!!!



Talk later and Regards,

Steve

[steve_qix]

in my experience (35+ years)

Holy moley, Steve, you are older than I thought.   Did you start with the broadcast transmitters when you were five or something?

Did you explain somewhere the distortion characteristics of the keep-alive supply you designed in the class H modulator circuit?  Is this just better than the 3 diode business?

Jon (three diodes in the Ranger but keep-alive supply in the class H/E)

Hi Jon !!  Thanks :-)  It's ALWAYS nice when someone thinks you are younger than you are :-)  I'm 53 years old (YIKES!!).

I started in professional radio at 18, while I was in college, and was chief of WORC when I was 21.

I always LOVED radio, and knew I would go into broadcasting someday, even as a little kid.

AND - lots of this is BECAUSE OF HAM RADIO, and the AM'ers in particular !!!!!  This is SUCH a GREAT hobby.  SWLing on an antique radio, listening to the AM guys on 160 and 75 meters, then parts from the dump (old TV sets) turned into a transmitter.... (license - what's that?).

Regards,

Steve

[WBear2GCR]

Ummmm....This discussion belongs in the Technical area...

The 3 diode limiter can doubtless be found with A) a search using the search function here or b) clicking on the AM Window icon up top and looking for it...


             _-_-


Dept of duh: http://www.amwindow.org/tech/htm/3diodeka.htm

[W7XXX]

Don, First I agree with your opinion on PEP and in my opinion it is a SSB term and has no business in reference to AM, however the FCC disagrees with me and has the power to take my hard earned license should I go by my theory that if a PEP meter measures my output it will show half what they claim by theory is my PEP. I mentioned it so any one interested in increasing their positive peaks would know how the FCC determines their power. I am still bitter that many so called AMers did nothing to fight the ARRL and FCC when they came up with the plan to discourage AM by reducing the power. The same when the deregulations of licensing was introduced to turn ham radio into CB. Many AMers saw this as a good thing, after all what has code to do with AM. Prehaps nothing, but is was a discipline that discouraged the CB types. In my opinion it is welfare licensing and another blow to AM.

Now, I agree with your theory on the single diode circuit, however in practicial use there are some discrepencies. I speak from experience using this circuit since 1999. I first experimented with the 3 diode and found it produces a ranchy signal that can be somewhat filtered by a following low pass filter. I then tried the single diode and resistor loading circuit and found that if the correct resistance is found, it works excellent with no hearable or seeable distortion. The theory is that since the modulating envelope reproduces the wave form of the modulating (voltage) signal accurately, there is no distortion. I find this to be the case as all attempts by those that call this method "a down and dirty way" to find distortion on my signal have failed. it is clean and never any splatter. I found this circuit in no way created a mismatch. I tried several taps on the UTC mod xfmr and always returned to the one that figured correctly mathematically.

The diode and resistor to ground as a protection device was used by many old homebrewers. In many cases this system wouldn't allow a reasonable amount of modulation because if the operator had a few voice frequencies that were of substantial amplitude over the others, this system saw overmodulation. The use of compression in some cases would help this but generally because of improper adjustments, led to distortion.

Since I am one that has a few voice freqs that are way higher in amplitude than others, it is difficult to get the majority of my voice freqs up to a reasonable level without overmodulation and splatter unless something is done to control negative peak clipping. I tried speech compression and it was hard to achieve the clean well rounded audio i desired at a high average level due to the amount of compression required to compress the offending few voice freqs. The single diode circuit bring all my voice freqs up to at least 100%, the average probably around 125% with the few that make the positive peaks jump to 140%.

I have tried low level speech clipping and didn't like it. It was too restricted.

Well, this is my experience that the single diode and resistor circuit can be made to work very well and is distortion free with no splatter if done right. One needs to bear in mind, modulator tubes need to be phased correctly and changing mics may change the phasing. Also the screen grid must be modulated with the plate. In the case of using a transformer with a screen tap, changes are you will never get this circuit to function correctly.

I would enjoy working you sometime and you can hear it for yourself and check it on your scope. I do commend you on your expertise that is superior to mine and appreciate your imformative post.

[w3jn]

Ummmm....This discussion belongs in the Technical area...

The 3 diode limiter can doubtless be found with A) a search using the search function here or b) clicking on the AM Window icon up top and looking for it...


             _-_-


Dept of duh: http://www.amwindow.org/tech/htm/3diodeka.htm

I *did* use the search function and it didn't come up 

[w3jn]

I have tried low level speech clipping and didn't like it. It was too restricted.

Sam, I don't think ANYONE here is advocating low-level clipping.  I think you will find that many of the modifications to Valiants, APaches, and the like advocated on this site involve bypassing any low-level clipping circuits.  I think it's generally accepted here that low-level clipping produces horrendous distortion.  Although you are against modification of these radios, this is definitely one mod that SHOULD be done.

Low-level processing (which you find objectionable) - yes.  Done correctly this will resolve the negative peak problem, is much easier and more convenient to adjust, and doesn't put undue stress on expensive high-level components.  As WB3HUZ pointed out it has many advantages, including the ability to tailor the audio for an empty band or crowded one.

Most of us wouldn't dream of putting an AM signal on the air without a scope.  It ensures quality control of the signal with the attendant courtesy towards other adjacent users on the band.

Steve (QIX) - I remember where I read that article; it was in ER a year or two ago and it was written by someone else but crediting you as the brains behind the concept   

73 John

[W7XXX]

John, You misunderstand what I was trying to say. The Ranger needs no mods in my opinion and experience. The Valiant is a whole different story. It needs lots of mods. Coupling caps up to .02, cathode bypass at 10 mfd, by pass the clipper, wire the 6146 modulators as triodes, beef up support on that flimsy chassis, change the 866's to 3B28's, etc. I never like the Valiant design. 3 6146's to do the job of 2, etc. Now the Apache and DX-100 also in my opinion need some basic audio mods, again coupling caps at .02, etc. Bear in mind I modifyed my old vintage rigs to match dynamic mics. The stock audio was ok for restricted audio using a xtal mic but hook up a dynamic and they were terrible in my opinion until modified.

My point was in my opinion based on 51 years of radio building, 3 years electronic trade school, former employment as parttime broadcast engineer at 3 AM radio stations (69-73), and running a successful audio repair shop for years, that changing coupling caps over .05 (.1 was mentioned) in a DX-100 for example doesn't improve the audio. .02 seems to be fine for well rounded audio. I was a little harsh in critizing those that do and feel like they have proved something, and for that I apoligize. I think the problem revolves around those who parrot what some very experienced and educated person has said. They don't understand why the enginneer made the mods and don't really care. If it is the popular thing they promote it. I respect the many of you here with more knowledge and experience than myself and even if my experiences disagree with yours, I still respect that your experiences are also valid because of the variables. Antenna building is a good example. You might build the same antenna as mine and it wouldn't be worth a rats rectum because of surrounding objects, location, or grounding methods. See my point? I never meant to talk down to anyone and didn't realize my opinions and experiences would be considered talking down. if folks look for faults in wording and dwell on it unstead of looking for the valid point, then trouble starts. The internet is famous for those who get their jollies by taking an opposing view point just to cause trouble. I enjoy a good debate and appreciate replies such as Don's where he explains in detail his stand on an issue.

[steve_qix]

Now, I agree with your theory on the single diode circuit, however in practicial use there are some discrepencies. I speak from experience using this circuit since 1999. I first experimented with the 3 diode and found it produces a ranchy signal that can be somewhat filtered by a following low pass filter. I then tried the single diode and resistor loading circuit and found that if the correct resistance is found, it works excellent with no hearable or seeable distortion. The theory is that since the modulating envelope reproduces the wave form of the modulating (voltage) signal accurately, there is no distortion. I find this to be the case as all attempts by those that call this method "a down and dirty way" to find distortion on my signal have failed. it is clean and never any splatter. I found this circuit in no way created a mismatch. I tried several taps on the UTC mod xfmr and always returned to the one that figured correctly mathematically.

You absolutely have distortion of the modulating signal if you are modifying its shape in any way (and/or if the modulator has its own distortion, which, unless it is VERY well designed, does).  Clippers in any form, including this one WILL distort the signal.  And, distortion WILL produce harmonics of the original waveform.  The single diode-resitor circuit will actually produce considerably more distortion than, say, the 3 diode circuit because the single diode circuit starts modifying the waveform earlier in the cycle.  The other (3-diode) method will only modify the most extreme peaks, which occur much less frequently, therefore less distortion.  

If you had a more raunchy signal with the 3 diode circuit, something was probably wrong.  Did you try the ultimate distortion test?  This would be to feed a triangle wave, probably 400Hz, into the transmitter and observe the result on an oscilloscope.   Very, very few amateur transmitters can accurately modulate a triangle wave.  If it can do 400hZ, then try a 200hZ and a 1kHz triangle.  Any bending of the straight sides of the triangle, or "rounding off" of the triangle is distortion.  If you really want to have fun, after the triangle test, try a square wave !

I have tried low level speech clipping and didn't like it. It was too restricted.

Hang on - low level clipping is *identical* to high level clipping IF THE MODULATOR HAS LOW DISTORTION.  Low level clipping has real advantages over high level clipping.  In broadcasting, no commercial transmitters use high level clippers.  All employ low level clippers, because the transmitters will handle the more complex, processed signals - and -  Low level clipping also has the distinct advantage of allowing the transmitter's audio filtering (PWM filters, etc.) to prevent excessive bandwidth.

I use low level clipping in my own transmitters.  The clipper circuit is DC coupled to the DC coupled modulator, and is ahead the anti-aliasing (audio low pass) filters, and of course ahead of the high level PWM filters which are part of all Pulse Width Modulators.  This allows me to control the audio bandwidth, regardless of the level of clipping (which is pretty low, because I like to have a high fidelity signal on the air).

Regards,

Steve

[W7XXX]

Steve, Let me rephase to say no distortion that effects audio quality or causes splatter. By theory changing the waveform causes distortion, but if one cannot hear the distortion, then it is a mute point. If no splatter occurs and I assure you it doesn't, then I accomplished my goal in a simpilier fashion. The idea of the 3 diode circuit was to achieve very high levels of postive peak modulation. The single diode will not go much beyond 140% whereas the 3 diode will go 300% by theory anyway.

i know several that have used the 3 diode with success, but it didn't work out for me and i achieved what I wanted with a single diode and saw no reason to mess around any further with the 3.

[WD8BIL]

(........, because I like to have a high fidelity signal on the air).

AHHH..... there's your problem Steve !

[steve_qix]

John, You misunderstand what I was trying to say. The Ranger needs no mods in my opinion and experience.

And that's what a lot of this is about - opinion.  In my opinion, the Ranger would need major and significant modifications if I were to use one as a transmitter.  But, of course, this is my opinion.  Doesn't necessarily make it "right" - just right for my goals.

You have to first state the goals before making any judgements.  The ranger, by measurement and test (I've owned several) is pretty terrible (sorry) with respect to accurate and good quality audio reproduction.  The goal (my goal, I should say):  accurate audio reproduction over the audible range.

If the goal is communications quality audio, the unmodified ranger is quite acceptable as commercial transmitters go.

I think you would have received considerably less flack with respect to most of your posts if the presentations were made from a personal opinion standpoint rather than an authoritarian (judgmental) standpoint.  Real, quantitative measurements are always useful, of course, because facts speak for themselves.

Regards,

Steve

[Steve - WB3HUZ]

PEP is figured by carrier power times 4 ... times percent positive modulation.
Example: 300 watts carrier x 4 = 1200 watts x 1.25 (125% +) = 1500 watts PEP.
This is how the FCC figures it whether we like it or not.

This is how the FCC figures it.

97.3 Definitions.

(b) The definitions of technical symbols used in this Part are:

(6) PEP (peak envelope power). The average power supplied to the antenna transmission line by a transmitter during one RF cycle at the crest of the modulation envelope taken under normal operating conditions.

[K6JEK]

In many years of experimenting here, I've become very fond of the 'passive bandpass filter' method of controlling audio bandwidth. Not many on here would agree with me on this but in my experiments it seems to do the job nicely. Nice rolloff on the extreme lows and the extreme highs, leaving the important part of the human voice intact.

Mack,   I'm interested in your experience with passive bandpass filters. Do you have a circuit you like which you'd care to share? I've been meaning to try some circuits myself but perhaps procrastination has saved me again. Maybe I can just learn from your experience.

Jon

PS. All my friends think my stock box Flex SDR-1000 sounds better than my Rangers but does not sound as good my class E (thanks, Steve). Bear in mind these are West Coast Sound evaluations.

[Steve - WB3HUZ]

PS. All my friends think my stock box Flex SDR-1000 sounds better than my Rangers but does not sound as good my class E (thanks, Steve). Bear in mind these are West Coast Sound evaluations.

That doesn't surprise me. A stock Ranger sounds OK, but a Flex can sound great. A few quick frequency and distortion measurements would tell you why.

[W7XXX]

Mack, I think you got it right ... my perspective is from an amateur radio communications viewpoint, whereas most here have the pro balanced line broadcast audio viewpoint. Audio quality as heard by the human ear and what one sees on the scope or what theory teachs us, may differ. An example is modulating over 100% positive and holding the negative at 100%. If done correctly the human ear can hear no distortion, but by the scope and theory distortion exist. The original waveform has been distorted, no argument, but the human ear can't hear it.

Part of my beef with the new wave of bass dominated AM is I have an inner ear disorder and can't hear many lower frequencies with any amplitude. When these bass dominated guys come into the group, I can't hear half what they are saying because they have mask their midrange and highs. If they practice well rounded audio, then I can hear them fine and they still have those mellow lows for those that can hear. For some reason it has become a fad here in the west to adjust their compressors for bass dominance. The solution is if I don't like it, get off the group which is what I have started doing. The problem arises when a few of us introduce a new frequency for AM ... here they come to dominate and start another large group.

Now the single diode circuit allows me to maintain my well rounded audio and bring up the amplitude of my weaker voice frequencies with no hearable distortion, eliminating the need for compression. I made this post because I received a PM from an OM seeking info on the single diode circuit. Not much info out there on it because many that have tried it, failed and also many hams just parrot theory and have never experimented with it, so they discredit this simple effective circuit based on BS.

Building the circuit and making it work ... talks ..........BS walks. My opinion.

[The Slab Bacon]

Sam,
       All bass with no presence rize sounds like crap, I concur with you on this one. You just agreed to what I said in a previous post. the audio needs to be well BALANCED to sound right. Too much bottom end without enough presence rize sounds like mush and is definately hard to understand at times. It takes a lot more than thundering bass to sound good. Its kinda like having turkey with mashed potatoes and stuffing, but no gravy. But with properly adjusted audio gear you can have it all, including a slice of pie to boot!!

                                               The Slab Bacon

[W7XXX]

Amen OM ... now we are on the same frequency ...